JPH0414283B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an optical surface roughness meter, and in particular to an optical surface roughness meter intended for non-contact and highly accurate measurement of the surface roughness of optical components. It is related to the meter.

[Background of the invention]

First, a conventional means for measuring surface roughness of optical components and the like will be explained.

Conventionally, stylus-type surface roughness meters have been used to measure the surface shape and roughness of objects to be measured, such as optical parts. There was a problem in that the surface of the object to be measured was scratched when viewed microscopically. As a countermeasure to prevent this, a method has been attempted in which the reflected light from the measurement surface and the reference surface is interfered without optical contact, and the unevenness of the measurement, that is, the surface shape, is measured from this interference fringe (Francis A. Jenkins et al.:
Fundamentals of Optics, Mc Gram Hill
Kogakusha, 1976).

A method of measuring the surface roughness of a measurement surface from such interference fringes and its problems will be explained using drawings.

Fig. 2 is a sectional view of the main part for explaining the conventional interference fringe measurement method, Fig. 3 is an interference fringe pattern observed by the measurement method according to Fig. 2, and Fig. 4 is the same as Fig. 3. Ideal brightness/dark temporal change diagram of such interference fringes, No. 5
The figure is a diagram showing the influence of external vibration on the brightness of the interference fringes shown in FIG. 3.

In FIG. 2, 9 is an object to be measured, 10 is a reference mirror fixed stage movable in the X direction, a reference mirror held by 10a, 5 is a half mirror, and the illumination light 3 is reflected by the half mirror 5. The object to be measured 9 and the reference mirror 10 are illuminated. The reflected light 4 from the measurement object 9 and the reference mirror 10 interfere with each other, and the half mirror 5
be monitored through the This interference fringe looks like, for example, FIG. 3. Let the light intensities at points A and B in the figure be I _A and I _B , respectively, and change the optical path length by slowly moving the reference mirror 10 in the X direction at a constant speed using a motor, _etc. , I _B are thought to ideally change with time t or displacement X as shown in FIG. That is, the interference fringes appear brightest at I _nio , and the interference fringes appear darkest at I _nio . However, in reality, due to external vibrations from the surroundings of the measurement optical system, the light intensity I of the interference fringes is
The waveform changes as shown by the solid line in the figure, and differs from the ideal waveform (double-dashed line in the figure). The frequency components of this external vibration cannot be ignored even when placed on the vibration isolation table of the measurement optical system, and vibrations of several tens of Hz or less cannot be ignored, and even if the measurement resolution is improved to 0.1 μm or less, reproducible measurements cannot be achieved. and the measurement surface resolution is 0.1.
The limit was about mm.

As means for changing the optical path length, in addition to the above-mentioned method of directly moving the reference mirror fixed stage 10a, there is a method of inserting a wedge-shaped prism into the optical path and moving this wedge-shaped prism in a direction perpendicular to the optical path using a motor or the like. It is also known (Special application 1987-
No. 168313). Using this method, the second
Compared to the method shown in the figure, although the amount of change in optical path length can be made extremely accurate, the problem remains that measurement accuracy cannot be avoided due to the influence of external vibrations.

In addition, in the conventional interference fringe measurement method, in order to measure the phase relationship _between two interference fringes, the _phase difference _δ was detected. However, I _nio (or
The accuracy of detecting the point I _nac ) is when ∂I/∂t (or ∂I/∂x) becomes 0 in its vicinity, so
The problem was that the detection accuracy was the worst. Therefore, the detection resolution of _δX (FIG. 4) had a limit of about 0.1 μm.

Furthermore, as a method for extracting the fundamental frequency component waveform from the noise-superimposed waveform (sampling data) as shown in Figure 5, there are methods such as fast Fourier transform, but since the calculation time is long, Even if this method was applied to measure the surface roughness of the object 9 to be measured, there was a problem in that it could not be carried out in real time.

[Purpose of the invention]

The present invention improves the problems of the conventional technology described above and provides an optical surface roughness meter that is capable of measuring surface roughness non-contact, highly accurately, and in real time without being affected by external vibrations. Its purpose is to provide the following.

[Summary of the invention]

In order to achieve the above object, the present invention provides a coherent light irradiation optical system that irradiates coherent light of a single wavelength perpendicularly to each of a reference mirror and a measurement object, and a coherent light irradiation optical system that irradiates a reference mirror and a measurement object with coherent light of a single wavelength. vibration imparting means for one-dimensionally vibrating in the optical axis direction at a frequency higher than the frequency; and the coherent light reflected from the reference mirror that is vibrated by the vibration imparting means and irradiated by the coherent light illumination optical system. a detection optical system that detects an optical image of interference fringes caused by interference with the reflected light from the measurement object irradiated by the light irradiation optical system; A multi-pixel image sensor that receives light, converts it into a signal with periodicity for each pixel, and outputs it, and a signal with periodicity that is output from each pixel of the multi-pixel image sensor as the reference mirror vibrates. extracting a fundamental frequency component signal, detecting a phase difference Δt between adjacent pixels in the multi-pixel image sensor for the extracted fundamental frequency component signal over the surface of the measurement object;
Calculation of the phase difference δt detected over the surface of the object to be measured based on the relational expression δt/T×λ/2 (T is the period of the signal of the fundamental frequency component, λ is the wavelength of the coherent light) This is an optical surface roughness meter characterized by comprising: arithmetic processing means for measuring the surface roughness of a measured object by subjecting the object to surface roughness.

More specifically, in order to avoid the influence of external vibrations, the reference mirror is moved one-dimensionally in the optical axis direction to change the optical path length at a frequency higher than the vibration frequency of the problematic external vibrations (mainly several tens of Hz or less). A vibration imparting means is provided to vibrate the reference mirror, and the fundamental frequency component of the periodic signal output from each pixel of the multi-pixel image sensor is significantly higher than the frequency component due to external vibration as the reference mirror vibrates. The objective is to be able to distinguish and detect the phase difference Δt between adjacent pixels of the image sensor. In addition, the arithmetic processing means detects the phase difference δt between pixels at locations where the rate of change of the fundamental frequency component signal extracted from the periodic signal output from each pixel of the multi-pixel image sensor is large. The reason is that it has made it possible to measure surface roughness with high precision. In addition, using a high-precision standard measurement object, the surface roughness
Do is measured, and then the surface roughness is measured using the object to be measured.
By measuring DM and correcting the measured surface roughness DM with the measured surface roughness Do, it is possible to measure the surface roughness with even higher accuracy.

[Embodiments of the invention]

Hereinafter, the present invention will be explained with reference to Examples.

FIG. 1 is a schematic diagram showing an optical surface roughness meter according to an embodiment of the present invention, and FIG. 6 is a signal processing circuit diagram of the output of the multi-pixel image sensor in FIG. 1.

In FIG. 1, 1 is an illumination optical system in which a filter is attached to a laser or an illumination lamp, 2 is an illumination lens system such as a collimator, 11 is a beam splitter, and 12-1 and 12-2 are objective lenses. , 14 is a piezo element related to high-speed driving means, which has a reference mirror 13 attached to its tip. By driving this piezo element 14 and reference mirror 13 at high speed in the optical axis direction (for example, 500 Hz, vibration 5 μm), This changes the optical path length. 22 is a multi-pixel image sensor such as an area sensor (CCD, CCPD, etc.) in which a large number of light-receiving elements are arranged one-dimensionally, a TV camera, etc. The signal processing circuit for the output of this multi-pixel image sensor 22 is as follows. 6th
This will be described later using figures. 20 is an intermediate lens, and 21 is a photographing lens. Note that the magnification of the objective lens 12-1 on the side of the reference mirror 13 can be made lower than the magnification of the objective lens 12-2 on the side of the measurement object 15, and these objective lenses 12-1, 12-2
Both the intermediate lens 20 and the photographing lens 21 can be composed of a plurality of lenses.

Reference numeral 16 is such that the object 15 to be measured is placed thereon, and the positioning and tilting of the object 15 can be adjusted in the height and lateral directions using an attached positioning and adjusting device (not shown). Stage 17 is
A drive table 19 on which the stage 16 is placed and fixed and is driven and fed by a motor 18 is a vibration isolation table on which the entire optical system is placed.

Next, a signal processing circuit for the output from the multi-pixel image sensor 22 will be explained using FIG. 6.

25 is an analog multiplexer that takes out the i-th (i=1, 2, 3, ...) element output and the i+1-th element output of the multi-pixel image sensor 22;
6-1 and 26-2 are both amplifiers (or buffers), 28-1 and 28-2 are both band-pass filters (BPF for short) (or high-pass filters (HPF for short)), 30-1 ,30
-2 is a reference value setting unit, 30-1 and 31-2 are comparators, 32 is a theoretical calculation circuit, 33 is a digital calculation circuit, 35 is a memory, 36 is a subtraction circuit, 37
is a DA converter.

A method for measuring the surface roughness of the object to be measured 15 using the optical surface roughness meter configured as described above will be explained.

As for the measurement procedure, first, using a measurement object that is close to a perfect mirror surface with very good surface roughness and shape accuracy, the calculation output D _O (details will be described later) due to distortion of the optical system is determined for the entire measurement range. Next, a calculation output _DM is obtained using the measurement object 15 that is actually desired to be measured. As shown _{in FIG .}
-D _O is subtracted, and the surface roughness of the measurement object 15 is thereby output digitally. DA this
The surface roughness is converted into an analog signal through a converter 37.
This is obtained as Vout.

The calculations for D _{O and} D _M are performed using the same procedure, but below is the measurement object 1 that you actually want to measure.
Regarding 5, the procedure for obtaining the above calculation output D _M is as follows:
This will be explained using drawings.

Fig. 7 is an output characteristic diagram schematically showing the output of interference fringes imaged on one element of the multi-pixel image sensor in Fig. 1, and Fig. 8 is an output characteristic diagram after passing through the bandpass filter in Fig. 6. Figure 9 is the 6th
Output diagram after passing through the comparator in Figure 10
The figure is a vibration waveform diagram of the piezo element in FIG. 6.

First, when the object to be measured 15 is placed on the stage 16, the positioning and adjusting device focuses the objective lens 12-2 and adjusts the tilt angle with respect to the optical axis.

Light emitted from an illumination light source 1 passes through an illumination lens system 2 and is split by a beam splitter 11 into transmitted light 3-1 and reflected light 3-2. The transmitted light 3-1 passes through the objective lens 12-1 and enters the reference mirror 13 attached to the tip of the piezo element 14, while the reflected light 3-1
2 enters the measurement object 15 through the objective lens 12-2. The reference mirror 13 is driven at high speed by a piezo element 14 in the optical axis direction.

The reflected lights from the reference mirror 13 and the measurement object 15 are reflected and transmitted by the beam splitter 11, respectively, and then interfere with each other. After passing through the intermediate lens 20 and the photographing lens 21, this light forms an image of the interference fringes on the multi-pixel image sensor 22.

In this case, assuming that the external vibration in question has a frequency of 1 Hz and an amplitude of 0.5 μm, the vibration speed is approximately 3 μm/
s, and the resulting frequency component of the interference light at a certain measurement point is approximately 10Hz, assuming that the wavelength λ of the illumination light is 0.6 μm, and even when the measurement optical system is stationary, the interference light will fluctuate to this extent. occurs. On the other hand, in this embodiment, when the reference mirror 13 is vibrated by driving the piezo element 14 in the optical axis direction with a sawtooth wave as shown in FIG. 10 at, for example, 500 Hz and an amplitude of 5 μm, the vibration speed is approximately 2.5mm/s,
The frequency component of the interference light is approximately 8KHz. Therefore,
As shown in FIG.
It is detected as a waveform having a fundamental frequency component of about 8KHz along with the vibration (optical path length change x or time t), and since the external vibration is about 10Hz, the period T of the output signal I of the interference light is Approximately 1/800
This will be detected as a change in .

To explain this using FIGS. 6, 8 and 9, the analog multiplexer 25 selects the i-th (i=1, 2,...) of the multi-pixel image sensor 22.
Take out the element output and the i+1th element output,
After passing through amplifiers 26-1 and 26-2, the amplifier output 27 is passed through bandpass filters 28-1 and 28-2.
8-2, when the fundamental frequency component signal is extracted from the periodic signal output from each element (pixel) as the reference mirror vibrates as shown in FIG. 7, the outputs 29-1, 29 -2 becomes Vi and Vi+1, respectively, as shown in FIG. Here, the DC component is removed, and the time integral value of the positive value and the time integral value of the negative value of the output Vi become equal. Output Vi from comparators 31-1 and 31-2,
Vi+1 is the reference value setting section 30-1, 30, respectively.
When compared with the reference value set to -2, rectangular waves CPi and CPi+1 as shown in FIG. 9 are obtained, respectively. The reference value of the comparator may generally be OV (that is, the point where ∂Vi/∂t, ∂Vi+1/∂t is maximum).

Next, the logic operation circuit 32 extracts a rectangular wave corresponding to the phase difference δt between the two rectangular waves CPi and CPi+1 in FIG. In Figure 9, the square waves CPi, CPi
Although δt is extracted at the rising edge of +1, it can also be extracted at the falling edge. If the rectangular wave δt and the reference clock pulse f _O are ANDed and counted by a counter, a digital output D (δ _t ) corresponding to δ _t can be obtained. Similarly, the period T of the rectangular waves CPi and CPi+1
Alternatively, if the rectangular wave corresponding to half period T/2 and the reference clock pulse f _O are ANDed and counted by a counter, a digital output corresponding to T or T/2 can be obtained. In the signal processing circuit according to FIG.
The digital output D(T/2) corresponds to 2.
Next, the digital arithmetic circuit 33 performs multiplication and division of λ/4×δt/T/2 (1). What is thus obtained is a step between the i-th point and the i+1-th point on the surface of the measurement object 15. However, λ is the wavelength of light. or It is also possible to obtain an averaged output. Although the above calculation is performed by the digital calculation circuit 33, it goes without saying that it can also be performed by using an analog calculation circuit.

When the calculation of equation (1) or (2) above is completed, the channel of the multiplexer 25 is sequentially switched to i+1 and i+2 by the calculation end signal (clock pulse) 34. On the other hand, the above calculation output is stored in the memory 35 by the calculation end signal 34.

When the above calculation and memory are completed for the entire measurement range of the multi-pixel image sensor 22, the feed table 17 is fed by the necessary amount by the motor 18, the same measurement is repeated, and the calculation output D _M is transferred to the measurement object 15.
is calculated for the entire measurement range of the surface of As mentioned above, the subtraction circuit 36 performs subtraction D _M -D _O with D _O obtained in advance for the object to be measured that is close to a perfect mirror surface, and this is passed to the DA converter 37 to convert the object 15 to the object to be measured.
The surface roughness of is obtained as an analog signal Vout.

In addition, in the signal processing circuit according to FIG. 6, the relationship between the measurement time of δ _t and T/2 and the calculation time of equations (1) and (2) above is as shown in FIG. Since the outward stroke of vibration No. 14 is the measurement time and the backward stroke is the calculation time, real-time measurement of surface roughness is possible. A specific example will be explained.

By using an objective lens with a high magnification of 100 times, a measurement surface resolution of 0.5 μm and a surface roughness measurement accuracy of 1 nm were obtained. Furthermore, by using a high-speed digital calculation circuit, it was possible to measure and calculate surface roughness within a measurement range of 2 mm in less than 4 seconds.

According to the embodiment described above, the reference mirror 13 is driven at high speed by the piezo element 14, and the signal processing circuit is made to perform the calculation of the above formula (1) (or formula (2)). The effect is that surface roughness can be measured non-contact, highly accurately, and in real time without being affected by vibrations. In addition, since the phase difference δt is detected where the change in the light intensity of the interference fringes is maximum, that is, where ∂Vi/∂t and ∂Vi+1/∂t are maximum, the measurement accuracy is extremely high. There is.
Furthermore, by setting the magnification of the objective lens 12-1 on the side of the reference mirror 13 to be lower than the magnification of the objective lens 12-2 on the side of the object to be measured 15 and lowering the numerical aperture NA, the surface roughness of the reference mirror 13 is slightly deteriorated. It also has the advantage that it can be treated as a surface that appears to be close to a perfect mirror surface, making it possible to measure surface roughness with high precision.

In this embodiment, the reference value of the comparator is set to the bandpass filters 28-1 and 28-2.
I tried to set it using the output OV, but the first
As shown in Figure 1, the peak value detection circuit etc.
Detect I _nac and I _nio and calculate the average value (I _nac + I _nio )/
2 may be used as the reference value.

Also, the location where the phase difference δ _t is detected is ∂Vi/∂t,
∂Vi+1/∂t does not necessarily have to be the maximum value, but may be in the vicinity thereof.

Furthermore, the magnifications of the objective lenses 12-1 and 12-2 may be the same;
It goes without saying that even if -1 and 12-2 are removed, interference fringes can be measured at low magnification.

Furthermore, as shown by the broken line in FIG.
If you insert the 4-plate 23 and analyzer (polarizing plate) 24,
By using a plane polarized laser light source, interference fringes with good contrast can be obtained even when the reflectance of the measurement object 15 and the reference mirror 13 are different.

Finally, in this example, the surface roughness was measured in real time, but no calculations were performed during the measurement, and the data of D(δ _t ) and D(T/2) were stored in the memory. It is also possible to set the values in advance, perform calculations all together after the measurement is completed, and output Vout.

Next, another embodiment will be described.

FIG. 12 is a schematic diagram showing the main parts of an optical surface roughness meter according to another embodiment of the present invention.

In the embodiment according to FIG. 1, the piezo element 14
However, in this embodiment, the reference mirror 6 is attached to the tuning fork 7 which is excited by the excitation coil 8, and the objective lens 12-1 is
The optical path length is changed by reflecting the light 3-1 that is incident on the light 3-1.

Even with this configuration, the same effects as the embodiment shown in FIG. 1 can be obtained.

〔Effect of the invention〕

As explained in detail above, according to the present invention, it is possible to realize an optical surface roughness meter that can measure surface roughness non-contact, highly accurately, and in real time without being affected by external vibrations. It has the effect that it can.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing an optical surface roughness meter according to an embodiment of the present invention, FIG. 2 is a sectional view of main parts for explaining a conventional interference fringe measurement method, and FIG. ,
An interference fringe pattern diagram observed by the measurement method according to FIG. 2,
FIG. 4 is a diagram of the ideal brightness and darkness of the interference fringes shown in FIG. 1
FIG. 7 is an output characteristic diagram schematically showing the output of interference fringes imaged on one element of the multi-pixel image sensor in FIG. 1, and FIG. 8 is a signal processing circuit diagram of the multi-pixel image sensor in the figure. , the output diagram after passing through the bandpass filter in FIG. 6, FIG. 9 is the output diagram after passing through the comparator in FIG.
0 is a vibration waveform diagram of the piezo element in FIG. 6, FIG. 11 is a waveform diagram showing another example of the reference value of the comparator, and FIG. 12 is an optical surface diagram according to another embodiment of the present invention. FIG. 2 is a schematic diagram showing the main parts of a roughness meter. 3-2... Reflected light, 6... Reference mirror, 7... Tuning fork, 8... Excitation coil, 12-1, 12-2...
Objective lens, 13...Reference mirror, 14...Piezo element, 15...Measurement object, 22...Multi-pixel image sensor, 25...Analog multiplexer, 26
-1, 26-2...Amplifier, 28-1, 28-2
...Band bus filter, 30-1, 30-2...
... Reference value setting section, 31-1, 31-2 ... Comparator, 32 ... Logical operation circuit, 33 ... Digital operation circuit, 35 ... Memory, 36 ... Subtraction circuit, 37 ... DA converter, δ _t ...Phase difference in changes in light intensity, T...period of light and darkness in light intensity.

Claims (1)

[Claims] 1. A coherent light irradiation optical system that irradiates coherent light of a single wavelength perpendicularly to each of a reference mirror and a measurement object;
a vibration imparting means for one-dimensionally vibrating the reference mirror in the optical axis direction at a frequency higher than the vibration frequency of the external vibration; a detection optical system that detects an optical image of interference fringes caused by interference between the reflected light from the reference mirror and the reflected light from the measurement object irradiated by the coherent light irradiation optical system; and the detection optical system. a multi-pixel image sensor that receives an optical image of interference fringes detected by a pixel, converts it into a signal having periodicity for each pixel, and outputs the signal; and as the reference mirror vibrates, each pixel of the multi-pixel image sensor The fundamental frequency component signal is extracted from the periodic signal output from the signal, and the phase difference δt between adjacent pixels in the multi-pixel image sensor for this extracted fundamental frequency component signal is calculated over the surface of the measured object. Regarding the phase difference δt detected over the surface of the object to be measured, the relational expression δt/T×λ/2 (T is the period of the signal of the fundamental frequency component, λ is the wavelength of the coherent light 1. An optical surface roughness meter characterized by comprising: arithmetic processing means for measuring the surface roughness of an object to be measured by performing arithmetic operations based on the following. 2. The optical surface roughness meter according to claim 1, wherein in the coherent light irradiation optical system and the detection optical system, objective lenses are installed in front of the reference mirror and the object to be measured, respectively. 3. Optical surface roughness according to claim 2, characterized in that the magnification of the objective lens installed in front of the objective lens is lower than the magnification of the objective lens installed in front of the object to be measured. Total. 4. The optical surface roughness according to claim 1, wherein the arithmetic processing means is configured to detect the phase difference δt at a location where the rate of change of the signal of the fundamental frequency component is large. Total.